@ArthurReutenauer indeed. I haven't seen any of the operas/plays/etc. of the story, but my grandfather would tell the story sometimes. Rather unfortunate to have the family name tied up in the bad side of a legend like that.
@PaulGessler If that’s any consolation, even William Tell was mythical, aparently.
@PaulGessler If you feel like it, we broadcast our version of William Tell later this month to cinemas all over the world, but the closest place to Milwaukee that does it seems to be Toronto ;-)
@daleif Last time I spoke to him, Morten wanted to be in charge of breqn and friends again… does anyone know if he'll be coming to TUG2015? Are you able to make it, Lars?
@WillRobertson I'm not going to attend. Come to think of it, last time I attended was in 2005. Conferences isn't really my thing. Most of it will be online afterwards anyway.
@daleif Well, I may agree with that but the fact is that if conferences didn’t exist, the community would lose a lot. I understand you don’t feel like attending yourself.
@daleif But personally, if I wasn’t going to conferences, I wouldn’t do any development.
@clemens Why defining printauthor and printitle? And not simply using the titlepage environment? Whay loading math packages and graphicx for a titlepage?
@daleif Oh, I’m having a different problem: if I didn’t meet fellow developers at conference, I would just hate all users and develop things for myself only :-)
@Johannes_B But if no-one ever tells a person that their code is crap, how can the make the crap go away. It has taken me over 10 years to get people here to write better LaTeX. Nowadays I see an increase in e.g. itemize + \item[(i)] because people don't know how to do lists and found that crap online
@Johannes_B And also one of the reasons I'm not a fan of generic templates. You always end up with too much stuff. And the end user does not know what half of it does due to poor documentation.
@yo' I didn’t know it was Czech children he rescued. I knew of the Kinderstransporte because there’s a commemorative sculpture in front of Liverpool Street station.
@daleif About the generic template issue, we’ve had a discussion with a user on this very chat not two days ago - he said he had a huge generic template and was adamant he would continue like that. See chat.stackexchange.com/transcript/message/22503443#22503443 and the discussion follows.
@ArthurReutenauer True, but in my experience most of the users I encounter are quite happy to learn to make their things better and not use some crappy solution they found through google and never gave an extra though about.
In most cases the user was not even aware that what he/she was doing was problematic.
Mostly whenever I edit book manuscripts or preprints I make a list of comments to the authors asking them to take them into account in their next project. And thus I see less problems with experienced users
@ArthurReutenauer Most people here know I can be quite harsh at times (though taken with a grain of salt). But then again they can just seek help elsewhere ;-)
@daleif The thing is, a beginner looking for a start, searches template, becuase this is the term. Finds something that advertises itself as high-quality. How would a beginner doubt the quality? A starter cannot judge the quality of code.
@ArthurReutenauer Well it is complicated. I was part of the departments IT section, but when the university merged all the IT sections into one unit I was asked to stay as an employee at the department as the IT unit would probably not allow be to do LaTeX support, which is very important to my department. In later years we have started publishing more and more books and thus I make sure they are publication ready. So I edit more LaTeX now a days than I actually write. (All hail Emacs)
@ChristianHupfer It was in my queue and i jut wanted to markup the code. That is the janitor in me. Suddenly, it was gone and i couldn't do anything more.
@daleif Interesting, I didn’t realise the city’s name had been officially changed. But back to the discussion about TeX, my point that when you interact with users you don’t know online, you really can’t say “your template is crap” even if it is. (You also can’t say that to your father face to face - I tried - but that’s a different issue. Besides, now I don’t even try any more, I just write crappy code on his behalf - lately by hacking one of the AMS classes.)
@ArthurReutenauer Of course, they cannot know your intentions behind it and may find it rude. But we still have a duty to inform people that that this and this is not good practise and should not be promoted.
@JosephWright The Å was invented in the 50s (I think) and the University where I work have always spelled it with a Aa which is the old time way of writing this special vowel. So to me it does not matter much.
@ArthurReutenauer True, but consider if we did not. Then also this site would be filled with bad solution that people would then take for granted that this is the proper way of solving this problem.
@clemens By looking at the code i was able to provide a redefinition as a quick fix. Now, the sectioning stuff is generated on the fly using the csname mechanism. I have to look at the code, just to understand what the doc is trying to tell me sometimes.
@Johannes_B I must say I always found scrguide hard to scan for relevant information. But writing manuals is not an easy task, anyway. (And neither a very pleasant one...)
@daleif No, we shouldn’t give up ;-) What I mean is that this serves internally, for “people in the know“ to recognise each other and filter out the contributions by uninformed users. This may sound elitist, but I think that on the contrary that’s what helps the community to improve on the very long term - and it also means that we have a duty to help other people improve their knowledge, and it’s failing to do that that would be elitist.
@ChristianHupfer no I assume the option not being known is just his distribution is too old, but as for what it actually does when it is there, I haven't a clue.
@DavidCarlisle, @ArthurReutenauer: We've solved the option error already in the comments. I referred to the xelatex issue there. Sorry, I was not precise enough
@ChristianHupfer Oh right. let me have a look at that ...
@ChristianHupfer There was a discussion fairly recently about \pdfminorversion in XeTeX (in the past few months) but I can’t remember what the outcome was.
@ArthurReutenauer ye, it has something to due with moving ice or melting ice during the last ice age. It formed a lot of the country side in Scandinavia
@Johannes_B It doesn’t need to be conscious. And I don’t think Socrates is to be taken literally on that one ;-)
@Johannes_B Has it never happened to you to read someone else’s answer to a question you already answered, and think “oh yes, that’s a good solution too, I can use that next time”? That’s the kind of thing I mean.
@ArthurReutenauer My point is, the people advertising their (bad) code, obviously don't know that it is bad. Sometimes i answer a question by ahcking something, whereas there is a clean interface i didn't know.
@Johannes_B Yes, but then we go back to the point I was making earlier: You have to educate users, but it’s a long-winded affair. And I would even argue that the main issue is not to make them realise that their code is bad, but that there is something to be learnt about the quality of the code.
@Johannes_B Any news from Martin about the uk,tug.org DNS entry?
@Johannes_B Sorry, meant to ask @JosephWright And yes, the DNS entry is correct, I don’t know why I had problems early (but the entry was definitely missing from where I was).
@Johannes_B You can’t go the other way round. If they don’t know there is something to learn, they can’t learn.
Hi, I asked a question concerning pgfplots yesterday, but nobody answered it, because it seems to be too specific. Can I restructure my question and repost it as a new question?
@Johannes_B Part of the problem is probably that there are so many ways to do anything in TeX, including very crappy ones. But newcomers may not realise that and look for a way to do something, which then becomes the way do to it.
@ArthurReutenauer Common sense would disagree. I had to learn how to read, later learn how to drive a car. People are studying to get a degree. You need to have some knowledge of a nuclear power plant in order to run it. Same with LaTeX, same with Word.
@ArthurReutenauer Agreed, ever searched for introduction LaTeX via google? Stuff comes up, i would never have dreamed of.
@Johannes_B I’ve heard people blame Microsoft for that (but you could make a valid point that Apple contributed to it a lot too).
@Johannes_B “Why” is irrelevant, although we can come back to it later if you want. If you agree that many people do not realise that LaTeX (or Word) have to be learnt, do you see my point then?
@DavidCarlisle It is not just a simple edit. My question was: "Plotting boxplots with pgfplots that have thick line at the mean value". I want to make it more broad. For example I can ask "How can I change the width of an specific line in pgfplots".
@ArthurReutenauer So that would be an issue with programs just work out of the box click play and the video plays. Hit a button in word, toggle italics and it is italic. Though not the way to do it. Hm.
Thanks. I have already done several edits on the original question: adding a minimal example and the picture of my desired output. So, I'm going for the new restructured broad question.
@Johannes_B We could work on the OP of tex.stackexchange.com/questions/253077/… (the question has already been mentioned several times here), because they seem a perfect illustration of the phenomenon I mentioned. No offence meant to that person, of course.
@clemens That’s hilarious (except when you’re that teacher, of course). There’s also the related phenomenon of blaming the messenger when they bring bad news.
@Ho1 it's your question, you can of course ask a new one if you wish, perhaps delete the old one if you no longer expect answers on it, I just wanted to mention that edits will automatically "re-advertise" the question.
@Ho1 You didn't add a "minimal example". Your code can't be compiled without extending it first as the preamble is missing. I tend to ignore such questions.
@Ho1 Ha, too bad. About right-to-left text, I understand that can be frustrating, but there are very few people who actually do anything to improve the situation, so there is little chance of it improving considerably any time soon.
Bidi implementation worked just fine in Context, but when it comes to LaTeX it is just a nightmare. Khaled described this as the "misuse of TeX scripts by LaTex", which seems to cause problems for the bidi algorithm that should run on the text itself.
I don't know if it is possible to fix this problem without rewriting LaTeX macros. So I gave up, because it is not a thing that I can do.
The quote was from my memory, I should say; But I think you can actually understand the situation. LaTeX macros was written without having bidi in mind.
I hope that, maybe in future versions of LaTex, for example LaTex3, we can draw the attention of LaTeX developers to the fact that we should have a unifed format, that supports multiple languages without too much effort.
Yes. Someone should write the code, fix the bugs and relase a working software distribution. But this needs too much effort. I'm mostly a LaTeX user, and not a LaTeX developer.
@UlrikeFischer I fixed the "minimal example". Now it just compile fine with xelatex.
@Ho1 we try:-) you have to remember though that latex was developed at a time when the only available tex engines were 8 bit and the only available fonts were 7 bit, so supporting any language other than English, even if it used the latin alphabet was a challenge. things are better now but still the state of the unicode support of the underlying engines is changing but that's good, people are developing it... So we'll get there, but it takes time
@Ho1 Well, too bad, but just so you know, the common practice in LaTeX is to mark any language switch, this has nothing to do with right-to-left scripts or bidi. In order to do fine typography you need to tag all content with a language.
You know what? We actualy HAVE a bidi-aware TeX engine called: Tex-e-Parsi. ctan.org/pkg/parsitex But it is old, and should be rewritten to match XeTeX.
@Ho1 saying something should be re-written doesn't make it happen. Almost all tex/latex development is done by volunteer developers in their spare time.
@ArthurReutenauer Yes, that's true. Even M$ Word does that. When you change your keyboard layout, it automatically sets the language of the words you write. But it is done transparently, so it is less annoying, and it is also invisible from the user (although it is shown in the status bar).
@Ho1 IMHO that's a problem with users in general: they want good results and transparency. The more we try to hide things from user, the more problems we get. Besides, I strongly believe that good typography is about fine tuning. When I submitted a paper to TUGboat, Karl and Barbara worked on every single sentence of mine in order to provide the most pleasant experience typography-wise.
I usually hear in TeX-aware math society that people use M$ Word because they are dumb, lazy, etc. But, at least for non-English rtl-languages, making good results needs too much effort.
@Ho1 How much did you pay for using LaTeX and friends or for getting help? The whole community does it for fun and to help others. Much of it is developed for own needs. We need more users that actually need and use RTL, because that is the only way to get feedback.
Really? Ok, forget about it! Can I say we agree that someone should fix the problems? As you said before: "as long as nobody writes code for LaTeX the situation won’t change."
@Ho1 Well one issue is that RTL does work for some people so we can't just change stuff: TeX--XeT has been around a long time (in XeTeX, pdfTeX, ...) while the Omega-based approach in LuaTeX also has a lot of history
@JosephWright While at the same time the TeX--XeTeX approach has major drawbacks if you want to produce complex documents - as I’m sure you’re aware. This is actually one the main things we want to fix with Jonathan, but we’ve got absolutely not time frame for the moment.
@ArthurReutenauer I was tempted to answer a recent mail on the XeTeX list saying that my big priority for XeTeX (and LuaTeX) would be something in this area, but I decided it would not be helpful
@JosephWright By the way, I of course didn’t tell you to send your latest request to the list to forget about it; on the contrary, I wanted to have it archived in a more permanent place than the chat, and I have taken it into account.
@ArthurReutenauer John Plaice talked at TUG2013 about a better RTL/TTB approach than the Omega-like one in LuaTeX, which sounded great if it can happen. If not, it would be very useful if XeTeX and LuaTeX at least did the same thing.
@Ho1 As you'll see from the discussion between @ArthurReutenauer and me, there is wider concern about RTL (Arthur is working on XeTeX, I do work on LaTeX itself)
@DavidCarlisle Which reminds me: RTL definitely on my 'quite important' list for TUG2015 expl3 discussions, at least in as far as having some kind of plan
@clemens At least one user wanted to have a \listofreactions (mychemistry.eu/2015/06/chemmacros-development) Still cannot comment. I think i also have mentioned that with the list before, though.
Great. Having a "best-practice" or a desired sitution can help having a good understanding of what needs to be done. Until now, the whole thing that we (I'm talking about myself) were concerned about was "How can I get Arabic/Persian text on my document, no matter how hard it is."
@ArthurReutenauer As I say, having some similar offering between XeTeX and LuaTeX really would be useful. As you will know, TeX--XeT is very limited and it's a headache in terms of potential LaTeX3 offerings in this area.
But I think it's now the time to think twice about the design. It's time to ask: "How to incorporate RTL languages in TeX, that works fine for most of the users"?
@Ho1 I don’t think that’s a fair assessment of the general situation. On the more specific issue of switching between a left-to-right and right-to-left language, why don’t use \XeTeXinterchartoks? -- if it’s XeTeX you use. If you have exactly one left-to-right language and one right-to-left language it should be easy enough.
I suggest to refer to some professional commercial (now open source) works like TeX-e-Parsi. It is considered a legacy TeX implementation now, but it does what Persian users want.
@ArthurReutenauer Right now, only Context have Unicode bidi implemented. Arthur Reutenauer said: "On the more specific issue of switching between a left-to-right and right-to-left language, why don’t use \XeTeXinterchartoks?" and I wanted to say that this does not fix the problem we are facing.
@Ho1 having bidi implemented in the engine would be better but it would seem so long as you had specified all the directional classes as xetex character classes, \XeTeXinterchartoks could be used to make a reasonable approximation to the unicode bidi algorithm that would cover most cases. Although if you are explicitly marking language change points most direction changes are explicitly marked anyway.
@Ho1 if all strong LTR characters are class 5 and all strong RTL characters are class 6 then you can specify that at the boundary between 5 and 6 you insert \beginR and at the boundary between 6 and 5 you insert \endR, so it's "just" a matter of reading in the UCD data files and then dealing with all the special cases
What I see as a best practice, can be deduced from what commercial/successful TeX implmenetations for various local languages have provied for sometime. Consider TeX-e-Parsi: You write your document, a combination of English and Persian language, the default is automatic-bidi and automatic font-selection. You only have to describe the paragraph direction and alignment, when it is needed.
This is totally different from the way things are done right now in XeTeX. Either with \lr, \rl, \latinfont or \persianfont in XePersian, or other means of changing language and direction, this is not the ultimate position I want to reach.
@JosephWright Yes, they did all kind of crazy stuff to make it as user-friendly as possible. The very fact that it’s not maintained any more should give an indication that it wasn’t sustainable in the long term.
Compiling the source without hyperref works perfectly well, but as soon as I use hyperref I get compilation errors.
%!TEX TS-program = xelatex
\documentclass[11pt, DIV=15, BCOR=10mm,parskip=half, headings=normal,openright,headsepline,
chapterprefix=off]{scrbook}
\usepackage{setspace}
\usepacka...
@ArthurReutenauer Whats? -> This is considered an ultimate goal in a bidi-aware Unicode TeX: "You write your document as a combination of LTR and RTL languages, the default is automatic-bidi and automatic font-selection. You only have to describe the paragraph direction and alignment, when you want to change the defaults. You don't have to change languages and font/scripts manually."
@Ho1 If you mean that it would be a good thing to have an option to do so, yes, of course it would be nice, but again nobody seems to be working on that.
@Ho1 In any case, since it deviates from the current TeX so wildly, it can’t possibly become the default behaviour in the foreseeable future.
@Ho1 I think this sentence says it all.
@Ho1 This has been fun, but if you don’t mind I”m going to stop contributing to this conversation now. It has gone on for long enough and I have to actually start doing the job I’m paid for at some point.
Why? English-language users have it right now. Don't they? I want to write Persian text, wherever I write English text right now, without having to write down multiple tags for direction/language/font swapping
@ArthurReutenauer Thanks. I'm subscribing to the mailing list. I'm sure it's better than chatting.
@Ho1 This chatroom is not the place to make any decisions about TeX development. If you have suggestions, please send them to some public mailing list or forum; if you have concrete questions, you can ask them on the TeX Stack Exchange. But this conversation has really become completely useless.
@Ho1 yes but the maintainers of the system have to care about other people as well. If you focus on one narrow concern things seem a lot easier.
@Ho1 you are concerned about two-language documents with exactly one LTR language and exactly one RTL language. That is a valid class of documents but if you design for that you will fail to cover the vast majority of multi-lingual documents.
@Ho1 but if even in a purely RTL document you need to put in an explicit marker every time you change language (so hyphenation works) then having some languages change direction as well is not such a big difference. It only seems bad to have to mark direction change if you are not marking language change and you can almost never do that for a wide class of documents.
@DavidCarlisle Considering the fact that manual selection of the direction causes several little problems that you won't notice until you carefully check every aspect of your document and manually fix it, I don't think the current situation with XePersian and bidi package is a good state to stay at.
@Ho1 no the state of bidi typesetting is unfortunately poor at present but the bidi algorithm (that is automatic detection of direction) isn't anywhere near the main problem, you first need to get the typesetting right when direction is explicitly marked, having tables columns coming out in the right order would be a good start.....
@Ho1 Not necessarily, but no point automatically detecting direction until other stuff fixed, it's a "nice-to-have" that can be layered on top but it isn't the first thing to work on (even if anyone was available to work on anything)
Compiling the source without hyperref works perfectly well, but as soon as I use hyperref I get compilation errors.
%!TEX TS-program = xelatex
\documentclass[11pt, DIV=15, BCOR=10mm,parskip=half, headings=normal,openright,headsepline,
chapterprefix=off]{scrbook}
\usepackage{setspace}
\usepacka...
